Frozen work units?

[Spezi]

Has anyone had any experience with work units feezing up?

I've got one dual core machine here that is really starting to annoy me as every time I check it it's either got one work unit frozen with one core idle or at times I'll even get both frozen up.

I've done the usual like install the latest BOINC client and even reformatted last night but it's still doing it.

Only way I can get it to restart is to restart the service or reboot the computer and then it will operate at 100% CPU for a while until one freezes up again.

Using XP Pro and I do not have that problem with my other machines.

 

[Adywebb]

Its sounding like some kind of hardware fault, possibly RAM?

What are the temps like when its at 100%?

Is it running at stock?

 

[Spezi]

Its sounding like some kind of hardware fault, possibly RAM?

What are the temps like when its at 100%?

Is it running at stock?

It's running at stock.

RAM is new and passes memtest.

What did happen though was the chipset fan died and I just replaced it with an aftermarket heatpipe style unit which has temps nice and low now.

Temps are under 40C when at full load.

 

[Adywebb]

I take it it was crunching fine before the chipset fan died?

Its definately down to a hardware problem of some kind, which when you are running the CPU at 100% something is locking up - something like this usually indicates overheating somewhere.

If it was fine before the chipset fan died, I'm wondering if damage has been caused when it failed.

 

[Spezi]

I take it it was crunching fine before the chipset fan died?

Its definately down to a hardware problem of some kind, which when you are running the CPU at 100% something is locking up - something like this usually indicates overheating somewhere.

If it was fine before the chipset fan died, I'm wondering if damage has been caused when it failed.

Yeah I'm beginning to wonder the same thing.

Temps are cooler than they ever were now.

 

[Adywebb]

You could try updating the chipset drivers?

Other than that the only thing I can suggest is trying another mobo

 

[Spezi]

You could try updating the chipset drivers?

Other than that the only thing I can suggest is trying another mobo

I have all the latest drivers from the ASUS site so I guess I'll try and find another board.

 

[Adywebb]

The only other thing is the PSU - are you able to test by doing a swop?

 

[Spezi]

The only other thing is the PSU - are you able to test by doing a swop?

Actually I just swapped the PSU for a new one because at first I thought the howling noise was the PSU fan but it turned out to be the Northbridge fan.

 

[Spezi]

Ady I"m a bit puzzled now as I changed the profile on that machine to crunch HPF and so far it's running on all cores without problems.

Starting to think my stuff doesn't like those Dengue work units.

I tried them when they first came out and had problems on several machines (also ceasing production until a reboot) which rectified itself by focusing only on Human Proteome work units.

A little while back I decided to give Dengue another shot and it seemed the machines were all OK except the one in question which I at the time attributed to the chipset cooling issue.

At this point I know the temps are great and can't help but figure if the Northbridge chip got fried there would be other signs of trouble and I don't see any other than frozen work units.

 

[Adywebb]

Thats weird, as there are no reports of that kind of problem with DDDT units - but at least you have that machine up and crunching again


Looks like you're just going to have to avoid them on that one.